NATIONAL PRIORITIES GROWING AMERICAN INEQUALITY S o u r c e s a n d Re m e d i e s B Y GA RY B U RT L ESS Over t he past t wo decades t he Unit ed St at es has experienced a st art ling increase in inequalit y. The incomes of poor Americans shrank and t hose of t he middle class st agnat ed while t he incomes of t he richest families cont inued t o grow. The well-being of families up and down t he income scale has increased over t he past five years, but t he average income of t he poorest Americans remains well below where it Gary Burt less is a senior fellow in t he Brookings Economic St udies program. This art icle is drawn from his chapt er in Set t ing Nat ional Priorit ies ( Brookings, fort hcoming).
was at t he end of t he 1970s. From t he end of World War II unt il t he 19 70s, t he percent age difference in average cash income bet ween well- t o-do and middle-class A merican f amilies generally declined ( see figure 1). In t he 198 0s t he gap began t o widen not iceably. Bet t er measurement of rich f amilies incomes account s for some of t he apparent jump in t he early 19 90s, but t he gap bet ween middle- and high-income families almost cert ainly increased aft er 19 92. The cash income dif f erence bet ween middle-income and poor f amilies f ollowed a similar t rend. Af t er narrowing f or several decades af t er World War II, largely because of increased wages and improved Social Securit y and welfare benefit s for t he poor, t he gap began widening in t he early 1 970 s. The figure suggest s t hat t he t rend in inequalit y has not been driven solely by worsening povert y among t he poor or by spect acular income gains among t he wealt hy. It has been produced by growing disparit ies bet ween Americans at every level of t he income ladder. Soaring inequalit y has not been confined t o t he Unit ed St at es. Rich nat ions around t he world have seen inequalit y grow since t he lat e 1 970 s. But t he jump in income inequalit y has been part icularly rapid in t he Unit ed St at es and it came on t op of a higher init ial level of inequalit y. Should We Care? Many Americans are not t erribly concerned about income inequalit y or about t he need for public policies t o t emper inequalit y. Alt hough public opinion polls find t hat large majorit ies of resident s in five European count ries and Japan believe t he government should guarant ee each cit izen a minimum st andard of liv ing, o nly abo ut a q uart er of Americans agree. By and large, Americans t end t o believe t hat people bear primary responsibilit y f or support ing t hemselves. U.S. cit izens are also more likely t o believe t heir societ y of f ers an equal opport unit y f or people who work hard t o get ahead. Given t hese views, why should Americans be concerned about mount ing inequalit y? One reason for concern is t hat growing income disparit ies may undermine Americans sense of social cohesion. Even if t hey are indif f erent about t he abst ract principle of economic equalit y, most Americans probably believe in t he ideals of polit ical and legal equalit y. But great er inequalit y has almost cert ainly produced wider discrepancies in polit ical influence and legal bargaining power. In 1 979 t he income of an American at t he 9 5t h percent ile of t he income dist ribut ion was t hree t imes Figure 1. Trend in Family Income Inequalit y: Rat io of Average Incomes of Well-t o-do and Middle-class Americans Income r at io of w ell- t o- do t o middle- class Income r at io of middle- class t o poor Well- t o- do Amer icans ar e t hose in t he t op 5 per cent of t he income dist r ibut ion. Middle- income Amer icans ar e t hose in t he middle 20 per cent. Poor Amer icans ar e t hose in t he bot t om 20 per cent. Sour ce: Aut hor s t abulat ions of Bur eau of t he Census P- 60 Repor t dat a. W I N T E R 1 9 9 9 3 2
t he median income and t hirt een t imes t he income of an American at t he 5 t h percent ile. By 1 996 an American at t he 95t h percent ile had an income almost four t imes t he median income and t went y-t hree t imes t he income of t he person at t he 5 t h percent ile. The growing income gap bet ween rich, middle-class, and poor and it s consequences for t he dist ribut ion of polit ical influence may cont ribut e t o Americans dwindling confi dence t hat t heir elect ed of f icials care very much about t he v iew s of ordinary cit izens. A c co rding t o polling exper t s Karly n Bowman and Everet t Ladd, in 1 9 6 0 only a quart er of U.S. respondent s agreed wit h t he st at ement I don t t hink public of f icials care much about what people like me t hink. By 19 96, t he share who agreed had climbed t o 60 percent. Inequalit y may also af f ect public healt h. Demographers and public healt h researchers hav e f ound mount ing t hough cont roversial evidence t hat great er inequalit y can boost mort alit y rat es and cont ribut e t o poor healt h. Count ries and communit ies wit h above-average inequalit y have higher mort alit y rat es t han count ries or communit ies wit h comparable incomes and povert y rat es but lower inequalit y. A c c o r d ing t o o n e p u b lic h ealt h r e sear c h er, lo w - inc o m e A m er ic ans hav e deat h rat es comparable t o t hose in Bangladesh, one of t he world s poorest count ries, even t hough absolut e incomes, average consumpt ion, and healt h care spending are much higher among America s poor t han t hey are in Bangladesh. The possible link bet ween public healt h and ine q ualit y m ay help ex p lain w h y t he Un it e d St at es, one of t he world s w ealt hiest count ries, does not have t he longest average life span or t he lowest infant mort alit y rat e. If t he benefi t s of U.S. income growt h aft er 1 979 had been more equally shared, t he average healt h and lif e spans of Americans, especially poor Americans, might have improved NATIONAL PRIORITIES MANY AMERICANS ARE NOT TERRIBLY CONCERNED ABOUT INCOME INEQUALITY OR ABOUT fast er t han t hey did. Def enders of American economic and polit ic al in st it ut ions correct ly point out t hat inequalit y plays a crucial role in creat ing incent ives for people t o improve t heir sit uat ions t hrough saving, hard work, and addit ional schooling. They argue t hat w age and income disparit ies must somet imes widen t o send correct signals t o people t o save more, work harder, change jobs, or get a bet t er educat ion. In t he long run, poor people might enjoy higher absolut e incomes in a societ y where income disparit ies are permit t ed t o widen t han one where law and social convent ion keep income diff erent ials small. According t o t his argument, widening inequalit y is in t he best long-t erm int erest of t he poor t hemselves. For poor people in t he Unit ed St at es, however, t he t heoret ical advant ages of great er inequalit y have proved elusive over t he past t wo decades. Their absolut e incomes have not improved; t hey have declined. Their absolut e incomes do not exceed t hose of low-income resident s in count ries wit h less inequalit y; t ypically t hey are lower t han t hose of people in a comparable posit ion in ot her rich count ries. The efficiency advant ages, if any, of growing U.S. inequalit y have not been enjoyed by t he poor, at least so f ar. They have flowed t o people much f urt her up t he income scale. Wh y Has In e q u alit y Increased? Resear c h er s o n inc o m e inequalit y agree on t wo key fact s. Great er family income inequalit y is closely connect ed t o wider disparit ies in worker pay disparit ies t hat in t urn are associat ed wit h rising pay premiums for educat ion, job experience, and occupat ional skills. In addit ion, shift s in family composit ion, specifically t he cont inuing growt h of singleparent families and t he shrinking fract ion of married-couple f amilies, have reinforced t he ef f ect s of widening wage inequalit y. W I N T E R 1 9 9 9 3 3
How much of t he increase in f amily income inequalit y is at t ribut able t o rising wage disparit ies? Bot h male and female workers saw hourly pay disparit ies increase over t he past t wo decades, t hough on average men saw t heir real earnings fall, while women got a raise. The hourly wage of workers at t he 1 0t h percent ile fell 16 percent bet ween 19 79 and 1 997. At t he upper end of t he pay ladder, wages at t he 90 t h percent ile rose 2 percent f or men and 2 4 percent f or women. Changes in annual earnings mirrored t his pat t ern. Workers at t he bot t om of t he pay scale saw t heir yearly labor incomes sink while workers at t he t op saw t heir annual pay increase. The gains were especially large among highly paid women. One way t o assess t he impact of rising wage disparit ies on overall income inequalit y is t o calculat e how much over all inequalit y would have c hang ed if w age dispar it ies had r em ained unchanged. My calculat ion, using a st andard st a- t ist ical measure of income inequalit y known as t he Gini coefficient, suggest s t hat if male annual earnings disparit ies had remained unchanged bet ween 19 79 and 1 996, personal income inequalit y would have increased about 7 2 percent of t he act ual jump. This means t hat t he increase in men s earnings inequalit y explains about 2 8 percent of t he overall increase in inequalit y. A similar calculat ion implies t hat despit e t he large increase in pay disparit ies among women, only about 5 percent of t he increase in income inequalit y can be explained by growing earnings disparit ies among women. We can combine t hese t wo calculat ions t o see what would have happened if male and f emale earnings inequalit y had bot h remained const ant after 1 979. This t hird set of calculat ions suggest s t hat t wot hirds of t he increase in personal income inequalit y would have occurred, even wit hout a change in pay disparit ies. An implicat ion of t his finding is t hat just one-t hird of t he increase in personal income inequalit y was due t o t he growt h of male and f emale earnings disparit ies. Most of t he growt h was due t o some ot her set of fact ors. One f act or was t he changing American household. In 1 979, 7 4 percent of adult s and children lived in married-couple households. By 1 996, t his share had fallen t o 65 percent. Inequalit y and t he incidence of povert y are much lower in marriedcouple households t han in single-adult households. If t he percent age of Americans living in marriedcouple f amilies had remained unchanged af t er 1 979, about one-fif t h of t he 1 979 9 6 jump in inequalit y would have been avoided. Anot her t rend has pushed up income disparit ies. Women who are married t o high-income husbands are increasingly likely t o hold year-round jobs and earn high incomes t hemselves. The increased correlat ion bet ween husbands and wives earnings has widened t he income gap bet ween affluent dual-earner families and t he rest of t he population. If t he husband-wif e earnings correlat ion had remained unchanged, about one-eight h of t he rise in overall inequalit y since 19 79 would have been avoided. In ot her words, roughly 13 percent of t he increase in income inequalit y can be t raced t o t he growing correlat ion bet ween husbands and wives earned income. Policy Response Though crit ics of U.S. social policy oft en overlook t he fact, policymakers have not st ood st ill in t he face of moment ous changes in t he income distribut ion. The direct ion of policy has shift ed not iceably since t he early 1 980s. The shif t began under President Reagan, who at t empt ed t o scale back and reorient welfare programs t arget ed on t he working-age poor. His goal was t o make t he programs less at t ract ive t o pot ent ial applicant s by cut t ing benefit s or making benefi t s harder t o get. One im port ant policy change, lat er reversed, was t o scale back payment s t o poor f amilies wit h a working adult. Reagan t hought w elf are benefi t s should be f ocused on t he nonworking poor. He expect ed working adult s t o support t hemselves. The st eep decline in hourly wages of low-skill workers made t his view increasingly unt enable. Measured in inflat ion-adjust ed dollars, t he minimum wage f ell more t han 3 0 percent over t he 198 0s, and wages paid t o unskilled young men fell almost as f ast. Few breadwinners can support families on wages of $5 or $6 an hour. Con gr ess and t he president responded by ref orming t ax policy t oward low-income f amilies and broadening eligibilit y f or publicly fi nanced healt h benefit s. The Tax Ref orm Act of 1 9 8 6 removed millions of low-income Americans f rom t he income t ax rolls and boost ed t he t ax rebat es low-income workers receive under t he Earned Income Tax Credit. The EITC was furt her liberalized in 1 990 and 199 3, great ly increasing t he credit s flowing t o low-income breadwinners and t heir children. Spending on t he credit increased eleven-f old in t he decade aft er 198 6, reaching more t han $21 billion by 199 6. The credit, payable t o breadwinners even if t hey owe no federal income t axes, has raised t he incomes of millions of f amilies wit h ext remely low earnings. The EITC is t he most dist inct ive American policy innovat ion on behalf of t he working poor, and sever- W I N T E R 1 9 9 9 3 4
NATIONAL PRIORITIES PUBLIC SUBSI- DIES TO THE WORKING POOR AND CUTS IN WELFARE BENE- FITS TO THE NON-WORKING POOR HAVE HELPED KEEP EMPLOYERS al European count ries may event ually adopt a variant of it. While most cash assist ance goes t o people who do not work, t he EITC goes only t o low-income people who do work. In 1997 t he credit provided as much as $3,6 56 t o a breadwinner wit h t wo or more dependent s. For a parent working full t ime in a minimum-wage job, t he EITC can increase net earnings nearly 4 0 percent. The idea behind t he credit is t o encourage work by increasing t he incomes available t o low-wage breadwinners who have dependent children. Inst ead of shrinking as a recipient s earnings grow, t he credit rises, at least up t o a limit. At low earnings levels t he credit increases by 34 or 4 0 f or each ext ra dollar earned. Most lab or eco nomist s who hav e examined t he credit conclude t hat it has cont ribut ed t o t he sudden and sizable increase in job holding among unmarried mot hers. Congress has also liberalized t he eligibilit y requirement s f or Medicaid healt h insurance t o include a broad populat ion of low-income children wit h working parent s. Unt il t he lat e 19 8 0s, working-age f amilies wit h children were usually eligible f or healt h prot ect ion only if t he f amilies were collect ing public assist ance. Children t ypically lost t heir eligibilit y f or f ree healt h insurance w h en t h e f am ily b r e ad w inn er r e t u r ne d t o w o r k. T h e Medicaid liberalizat ions of t he lat e 1 9 8 0 s and early 1 990 s meant t hat many children were enrolled in t he program even if t heir parent s had m o dest earnings and were not collect ing public assist ance. Some st at e government s have est ablished new programs t o provide subsidized healt h insurance t o members of w o rking -poo r f am ilies, including t he adult breadwinners. Congress passed legislat ion in 19 97 offering st at es generous f ederal subsidies t o est ablish or enlarge healt h insurance programs for t he working poor and near-poor. As U.S. policy has expanded t ax and healt h benefit s for t he working poor, st at e and federal policymakers have slashed cash assist ance t o t he nonworking poor. General assist ance, which provides cash aid t o childless adult s, has been scaled back or eliminat ed in several st at es. Aid t o Families wit h Dependent Children was eliminat ed in 199 6 and replaced wit h Temporary Assist ance t o Needy Families (TANF). The new federal program pressures all st at es t o curt ail cash benefit s t o poor parent s who are capable of working. The head of each family on welf are is required t o work wit hin t wo years af t er assist ance payment s begin. Work-hour requirement s are st ringent, and st at es face increasingly harsh penalt ies f or failing t o meet t hem. The law st ipulat es t hat t he great majorit y of f amilies may receive benefit s for no longer t han five years and permit s st at es t o impose even short er t ime limit s. Over a dozen st at es have already done so. The new welf are law and t he new st at e welf are policies t hat preceded it helped produce an unprecedent ed drop in t he nat ion s child welf are rolls. Since peaking in 1 994, t he number of f amilies collect ing public assist ance f or children has dropped more t han 2 million, or 4 0 percent. In sum, U.S. policy has become much less generous t o t he nonworking ( but working-age) poor, while it has become much more generous t o working-poor adult s wit h children. For many low-wage breadw inners wit h children, t he recent policy changes t he increased generosit y of t he EITC, Medicaid, st at e- support ed healt h plans, and child care subsidies hav e offset t he loss of pot ent ial earning s du e t o shrinking hourly wages. The r ef orms are hav ing ot her economic ef fect s. Poor breadwinners wit h children have been induced t o ent er t he work f orce and st ay t h er e. Th eir e n t r y c o n- t ribut es t o t he downward pressure on t he wages of t he least skilled. In effect, public subsidies t o t he working poor and cut s in welfare benefit s t o t he nonworking poor have helped keep employers cost s low and t hus helped fuel employers creat ion of poorly paid jobs. W I N T E R 1 9 9 9 3 5
Fut ure Direct ions U.S. policies t oward low-income, working-age f amilies are not so callous t hat st ruggling families have been left wholly on t heir own t o cope wit h declining wages. But t hey are not so generous t hat poor, working-age Americans have shared equally in t he prosperit y of t he past t wo decades. Dif f erent policies, such as t hose adopt ed in West ern Europe, would have yielded dif f erent result s. Some dif ferences, including lower povert y rat es and higher wages, make West ern Europe a more pleasant place t o live, especially for t he poor. But ot hers, including high unemployment, are unwelcome. It is not obvious t hat most Americans, even liberals, would pref er t he European approach or approve t he policies needed t o achieve it. While t he current U.S. policy mix broadly reflect s t he preferences of U.S. vot ers, it is haphazard and f ails t o reach some of t hose who most need help.two new policies could aid working-age people who have suf f ered t he worst cut s in hourly pay. The first would assure some of t he long-t erm unemployed a job at a modest wage. The second would make work subsidies more unif ormly available and would provide t hem in a form t hat most vot ers approve. Because public assist ance t o t he nonworking but able-bodied poor is being drast ically curt ailed, it makes sense t o assure at least some poor adult s t hat t hey will be able t o fi nd jobs at a modest wage, however bad t he local job market. In some cases t his may involve creat ing publicly subsidized jobs t hat pay a lit t le less t han t he minimum wage. It seems part icularly import ant t o ext end t his offer t o parent s who face t he loss of cash public assist ance. If vot ers and policymakers want unskilled parent s t o begin support ing t hemselves t hrough jobs, t hey should assure t hese parent s t hat some jobs will be available, at least event ually, even when unemployment is high. For poorly paid breadwinners, it is essent ial t o improve t he rewards from working. One possibilit y is t o make a basic package of subsidized healt h insurance available t o all children and young adult s. Many Americans regard healt h insurance for children as a fair and accept able way t o help t hose in need. Most healt h insurance for children is eit her publicly subsidized t hrough Medicaid or privat ely provided t hrough employer healt h plans. When insurance is financed by employers, most of t he cost t o employers shows up as lower money wages paid t o workers. By publicly assuming some or all of t he cost of paying for a basic healt h package for children, we could push employers t o boost t he wages t hey pay t o insured workers who have child dependent s. Such a move would have a great er impact on t he pay of low-wage workers, for whom healt h insurance represent s a big f ract ion of compensat ion, t han on t he pay of high-wage workers. About 1 5 percent of all children ( and nearly a quart er of poor children) have no healt h insurance. For t hese children and t heir working parent s, publicly subsidized child healt h insurance would direct - ly improve well-being and reduce out -of-pocket spending on medical care. It would also great ly increase t he reward t o work. Parent s who do not work qualify for free medical insurance f or t hemselves and for t heir children under Medicaid. Some lose t his insurance when t hey accept a job t hat pays modest but above-povert y-level wages. A public healt h insurance package f or all children would reduce or eliminat e t his penalt y f or accept - ing a job. American economic progress over t he past t w o decades has been quit e unev en. Families and workers at t he t op of t he economic ladder have enjoyed rising incomes. Families in t he middle have made much smaller income gains. Workers at t he bot t om hav e suf f ered a sharp erosion in t heir relat ive income posit ion. For some low-income workers, new public policies have helped offset t he loss of wages wit h larger earnings supplement s and bet t er healt h insurance. But m any low-wage workers hav e not benefi t ed f rom t hese policies. Humane public policy should t ry t o assure t hat t he most vulnerable Americans share at least modest ly in t he nat ion s prosperit y. W I N T E R 1 9 9 9 3 6